I have some schematics for old Neve desks here, all discrete and transformer coupled (35ish yrs old design this), sound lovely if a little rock and roll 🙂 I also have some more modern designs (V series Neve) NE5532s everywhere and lots of electrolytic caps (that fail regularly).
Who can host pics if I photo some schematics.
Who can host pics if I photo some schematics.
Re: Re: Re: learn from recording consoles?
True for near unity gain operation (AKA very high feedback) where the large common-mode component is manifest as a small increase in second order distortion.
Hardly significant, particularly as the gains in THD+N in the inverting case are likely to be swamped by the increase in noise to which this arrangement is subject.
jcx said:
True for near unity gain operation (AKA very high feedback) where the large common-mode component is manifest as a small increase in second order distortion.
Hardly significant, particularly as the gains in THD+N in the inverting case are likely to be swamped by the increase in noise to which this arrangement is subject.
Hi zenmasterbrian,
Preamplifiers for home do go this route. Many run rails of 25 to 35 VDC with no op amps in sight.
-Chris
Read more, then listen more. J Fet inputs have lower transconductance and can be more linear than BJT's. Input impedance is an issue for noise performance and that may make the decision. Most studio environments are low impedance as you mentioned in a round about way.
This will never happen in a real recording console. Heat generation is already a problem. In a recording console this would be a very poor choice. Many recording rooms may use one or two industrial airconditioning units already. The power supplies, amps and computers are often mounted in closets or rooms with air seals between.
Preamplifiers for home do go this route. Many run rails of 25 to 35 VDC with no op amps in sight.
-Chris
imix500, Wavebourn, kubeek, and Puggie,
I learn a tremendous amount from studying schematics, from every era.
If you email them to me, I will redistribute them to whoever is interested.
brian.zenmaster@gmail.com
Getting the full sized image is really helpful.
Also, we can post up to 1Meg here:
http://groups.google.com/group/audioex_preamp_source?hl=en
If they are too big, we can reference them and put my email address for those who are interested.
FWIW, your posts jog my memory, NE5532 are class A op amps for audio. I think they draw 20mA quiesent? There is higher spec stuff on the market now.
I learn a tremendous amount from studying schematics, from every era.
If you email them to me, I will redistribute them to whoever is interested.
brian.zenmaster@gmail.com
Getting the full sized image is really helpful.
Also, we can post up to 1Meg here:
http://groups.google.com/group/audioex_preamp_source?hl=en
If they are too big, we can reference them and put my email address for those who are interested.
FWIW, your posts jog my memory, NE5532 are class A op amps for audio. I think they draw 20mA quiesent? There is higher spec stuff on the market now.
<<
True for near unity gain operation (AKA very high feedback) where the large common-mode component is manifest as a small increase in second order distortion.
Hardly significant, particularly as the gains in THD+N in the inverting case are likely to be swamped by the increase in noise to which this arrangement is subject.
>>
Both noise and distortion do need to be analyzed. Both are complex.
Do you have analysis that shows inverting mode noise to be higher? Most phono stages use noninverting. But some use inverting. The noise coloration is different, lower freq for inverting.
Most line level functions are near unity gain. But since to get a fad to zero gain control, you really need noninverting, maybe best to do that where there is significant gain.
Many active filter examples use unity gain noninverting Salen Key.
I think better to look at the more obscure inverting Salen Key. See if it can be used.
True for near unity gain operation (AKA very high feedback) where the large common-mode component is manifest as a small increase in second order distortion.
Hardly significant, particularly as the gains in THD+N in the inverting case are likely to be swamped by the increase in noise to which this arrangement is subject.
>>
Both noise and distortion do need to be analyzed. Both are complex.
Do you have analysis that shows inverting mode noise to be higher? Most phono stages use noninverting. But some use inverting. The noise coloration is different, lower freq for inverting.
Most line level functions are near unity gain. But since to get a fad to zero gain control, you really need noninverting, maybe best to do that where there is significant gain.
Many active filter examples use unity gain noninverting Salen Key.
I think better to look at the more obscure inverting Salen Key. See if it can be used.
<<
Read more, then listen more. J Fet inputs have lower transconductance and can be more linear than BJT's. Input impedance is an issue for noise performance and that may make the decision. Most studio environments are low impedance as you mentioned in a round about way.
>>
The argument about JFETs producing square law distortion comes from their characteristic curves. But is that really how it pans out in an opamp circuit?
Every attempt is made to keep the current through the input diffpair from changing. But it still changes a little bit.
So, since the FET has lower transconductance than the BJT, the FET's Vgs will be changing more than the BJT's Vbe.
These amount to errors in differencing, and they are the major distortion source in opamp circuits with a large common mode signal. Meaning noninverting, large signal, and near unity gain.
So I doubt any argument that the FET input opamp is lower distortion. Is it square law? Not really.
I threw the JFET input argument out here because it occurs on two other threads and you see it many places.
D. Self and G. R. Slone reject it.
What Slone likes is mirroring. That is he provides a second upside down complementary diffpair. This raises 2nd order distortion just a bit, but the high frequeny and fairly high 3rd order distortion goes way way down.
His book is very interesting.
I'm looking forwards to seeing recording console schematics. I learn a tremendous amount from studying schematics, and considering what the alternatives might have been.
<<
"I would still say there could be advantages in going discrete:
1. Able to run hotter. Bigger parts, heat sinks, lower impedances, higher bias at all stages.
2. Could use higher voltages. "
This will never happen in a real recording console. Heat generation is already a problem. In a recording console this would be a very poor choice. Many recording rooms may use one or two industrial airconditioning units already. The power supplies, amps and computers are often mounted in closets or rooms with air seals between.
>>
Most home audio preamp stuff does not draw that much power per gain unit.
Of course some stuff is designed for batteries, and is real lean.
Monolithic circuits favor this. They have much lower interconnect capacitance to drive.
But I have read that the highest SNR 32 channel consoles draw 2kW. These must be high current circuits.
I believe LM394, one of the favorite NPN diffpairs, can be run at higher current than you would a monolithic opamps diffpair.
This can help with noise and slew rate.
You can have bigger output drive capability with discrete too. You can have a bigger second stage, which reduces its sensitivity to loading.
These power supplies in the closets, for the recording consoles; how many voltage rails, what are they, are they linear suppies or SMPS?
<<
Preamplifiers for home do go this route. Many run rails of 25 to 35 VDC with no op amps in sight.
>>
Are you describing things that have just one pair of +/- rails? I've thought about circuits that would use two pairs. That was part of the motivation for this thread, hoping to find some existing examples.
You say "no op amps in sight". Do you mean no monolithic opamps, or no opamp circuits?
Are you refering to circuits that do the feedback through the emitter node? Are you refering to single ended circuits?
Of course, in the name of subjectivism people do all sorts of things.
But I still love to study all schematics. Got any?
ZMB
thread reopened!
http://groups.google.com/group/audioex_power_supplies?hl=en
Read more, then listen more. J Fet inputs have lower transconductance and can be more linear than BJT's. Input impedance is an issue for noise performance and that may make the decision. Most studio environments are low impedance as you mentioned in a round about way.
>>
The argument about JFETs producing square law distortion comes from their characteristic curves. But is that really how it pans out in an opamp circuit?
Every attempt is made to keep the current through the input diffpair from changing. But it still changes a little bit.
So, since the FET has lower transconductance than the BJT, the FET's Vgs will be changing more than the BJT's Vbe.
These amount to errors in differencing, and they are the major distortion source in opamp circuits with a large common mode signal. Meaning noninverting, large signal, and near unity gain.
So I doubt any argument that the FET input opamp is lower distortion. Is it square law? Not really.
I threw the JFET input argument out here because it occurs on two other threads and you see it many places.
D. Self and G. R. Slone reject it.
What Slone likes is mirroring. That is he provides a second upside down complementary diffpair. This raises 2nd order distortion just a bit, but the high frequeny and fairly high 3rd order distortion goes way way down.
His book is very interesting.
I'm looking forwards to seeing recording console schematics. I learn a tremendous amount from studying schematics, and considering what the alternatives might have been.
<<
"I would still say there could be advantages in going discrete:
1. Able to run hotter. Bigger parts, heat sinks, lower impedances, higher bias at all stages.
2. Could use higher voltages. "
This will never happen in a real recording console. Heat generation is already a problem. In a recording console this would be a very poor choice. Many recording rooms may use one or two industrial airconditioning units already. The power supplies, amps and computers are often mounted in closets or rooms with air seals between.
>>
Most home audio preamp stuff does not draw that much power per gain unit.
Of course some stuff is designed for batteries, and is real lean.
Monolithic circuits favor this. They have much lower interconnect capacitance to drive.
But I have read that the highest SNR 32 channel consoles draw 2kW. These must be high current circuits.
I believe LM394, one of the favorite NPN diffpairs, can be run at higher current than you would a monolithic opamps diffpair.
This can help with noise and slew rate.
You can have bigger output drive capability with discrete too. You can have a bigger second stage, which reduces its sensitivity to loading.
These power supplies in the closets, for the recording consoles; how many voltage rails, what are they, are they linear suppies or SMPS?
<<
Preamplifiers for home do go this route. Many run rails of 25 to 35 VDC with no op amps in sight.
>>
Are you describing things that have just one pair of +/- rails? I've thought about circuits that would use two pairs. That was part of the motivation for this thread, hoping to find some existing examples.
You say "no op amps in sight". Do you mean no monolithic opamps, or no opamp circuits?
Are you refering to circuits that do the feedback through the emitter node? Are you refering to single ended circuits?
Of course, in the name of subjectivism people do all sorts of things.
But I still love to study all schematics. Got any?
ZMB
thread reopened!
http://groups.google.com/group/audioex_power_supplies?hl=en
What does anatech mean?
As I know, ANA means Automatic vector Network Analyzer.
This has been something for microwave and radio freq, and I have much experience with such. Like HP8510.
But I believe a company in Oregon makes one for audio?
Is that how you mean it?
As I know, ANA means Automatic vector Network Analyzer.
This has been something for microwave and radio freq, and I have much experience with such. Like HP8510.
But I believe a company in Oregon makes one for audio?
Is that how you mean it?
Hi Cal,
Right you are! Analog Technician...
Hey, I had to think of a name fast since the ones I wanted were gone already. I found out later there is a company of the same name.
Zenmasterbrian. HP is my fav test gear. My better stuff is HP.
-Chris
Right you are! Analog Technician...
Hey, I had to think of a name fast since the ones I wanted were gone already. I found out later there is a company of the same name.
Zenmasterbrian. HP is my fav test gear. My better stuff is HP.
-Chris
Hello Brian,
I am most often found lurking in the speaker section but the job does demand that I also check elsewhere.
I am most often found lurking in the speaker section but the job does demand that I also check elsewhere.
Hi zenmasterbrian,
BTW, the total tail current will not change at all for the sake of argument. It just seesaws back and forth. The gate is a much higher impedance at low audio frequencies and therefore will show a higher signal on it's gates. Not a problem at all in op amps or discrete circuits.
You should study MikeB's Symasym thread. It's rather long though.
-Chris
Exactly. A fet seems to have more linear range, a lower slope in a diff pair.
BTW, the total tail current will not change at all for the sake of argument. It just seesaws back and forth. The gate is a much higher impedance at low audio frequencies and therefore will show a higher signal on it's gates. Not a problem at all in op amps or discrete circuits.
Not as big a problem as you may think. The TL071 series exists after all, and many others.
Different strokes ........ Mr. Curl seems to like them. So do I and many others. Could have been the technology of the time for Mr. Self and Mr Slone. I wasn't there, so I have no idea why they don't like them.
My personal least favorite. This topology did allow for higher tail currents so that DC offset voltage could be kept down. I like fets even more for this in a straight differential. Cascoded of course!
I don't know. I do know that standard Neve and SSL consoles throw off quite enough heat already. Heat is a major enemy of reliability and noise. The use of high currents in RF is needed to charge and discharge capacitances. For audio you really need to reduce heat.
Bigger die and package = better dissipation. So yes - on average. There may be an op amp out there that runs hot.
They are increasingly SMPS due to heat. They are whatever heat producing things they can take out of the room. Some are hot, hot linear types.
Discrete circuits. Simpler that most op amps, or more complicated. Most do not require separate supplies as they operate in class A typically.
Not in particular. I don't like those, but that's just me.
What you have to do is limit your area of study so you can focus. Otherwise you get lost in a mountain of information. I love this field because I'm always learning.
You should study MikeB's Symasym thread. It's rather long though.
-Chris
I second pinkmouse, unless you're talking top-tier designs (which I haven't seen) all the consoles in my experience have been textbook examples of utilitarian audio design. Get it done reasonably and hopefully reliably, leave the hotrodding for Focusrite and Neve. Oh, and make it heavy. Metalwork leaves an impression of quality. 🙂
Hi rdf,
You are sooooo right. Same goes for outboard gear.
Do you know what they use for IC's? Beldin. Most of those snakes are a million years old too!
Example. You have a 24 track deck. How many channels is that back and forth? 72. Per machine. Balanced. Most use two machines. You are unlikely to run something like Canaire unless it's a mic line or some such. The quality doesn't suffer unless there is some defect with the cable or connector.
I didn't even mention the patch panel(s)
-Chris
You are sooooo right. Same goes for outboard gear.
Do you know what they use for IC's? Beldin. Most of those snakes are a million years old too!
Example. You have a 24 track deck. How many channels is that back and forth? 72. Per machine. Balanced. Most use two machines. You are unlikely to run something like Canaire unless it's a mic line or some such. The quality doesn't suffer unless there is some defect with the cable or connector.
I didn't even mention the patch panel(s)
-Chris
pinkmouse said:
That's why all tube consoletes are still produced. And they give light!
BTW
I can't understand why 741s are used whereas something way better, like tl07x is cheap as dirt. NE5534s aren't expensive too and are better than tl07x.
😕
And heat in the cold, cold north.
Seriously. These days tubes are found in outboard gear and guitar amps. That's about it.
-Chris
Seriously. These days tubes are found in outboard gear and guitar amps. That's about it.
-Chris
Yup, we use almost exlusively Wireworks cable for trunks at work, it's all Belden. We have a lot of Whirlwind W stuff, but it doesn't go out very often anymore. Wireworks did at one time use this wonderful,soft,flacid Mugami cable on thier W series, insanely expensive stuff.
Most new large frame consoles have switchers now. It's not so much heat, it's weight thats important. Lifting a 150lb PSU rack (and trucking one) is much nicer than a 450lb PSU rack in Cadac terms. But yeah, Midas, Digico, Cadac, Yamaha, all switchers.
I think the J-Type Cadac runs at +/- 27, +15, and +48, I'll have to double check at work tomorrow.
I'm working on those schematics. I just have to steal the service manuals away from our console guy long enough to copy them.
Hi imix500,
I was talking about the SSL consoles that use two computers. Fairly large. You sure as heck don't road those unless they are permanently mounted in a trailer. Weight is not an issue with those. The show consoles are another story again.
-Chris
I was talking about the SSL consoles that use two computers. Fairly large. You sure as heck don't road those unless they are permanently mounted in a trailer. Weight is not an issue with those. The show consoles are another story again.
Careful. You'd better ask. 😉 The techs I know would kill anyone taking a manual without asking. Rightfully so. Aren't those on CD by now?
-Chris
Hey Chris,
Ah yes, learned on a SSL (7000 I think?)in college back when I thought I wanted to be a studio rat.🙄
Ah yes, learned on a SSL (7000 I think?)in college back when I thought I wanted to be a studio rat.🙄
Heh, no worries. I always ask. Besides, I'm pretty well in with the console guys. Built up a Digico D5 for Rod Stewarts tour today with them. Just gotta find time when they don't need the manual to service a Cadac, which isn't very often.
- Status
- Not open for further replies.